rocAL_pybind.amd.rocal.plugin.tf.ROCALIterator Class Reference

rocAL_pybind.amd.rocal.plugin.tf.ROCALIterator Class Reference#

rocAL: rocAL_pybind.amd.rocal.plugin.tf.ROCALIterator Class Reference
rocAL_pybind.amd.rocal.plugin.tf.ROCALIterator Class Reference

ROCAL iterator for detection and classification tasks for TF reader. More...

Inheritance diagram for rocAL_pybind.amd.rocal.plugin.tf.ROCALIterator:
rocAL_pybind.amd.rocal.plugin.tf.ROCALGenericIteratorDetection

Public Member Functions

def __init__ (self, pipelines, size=0, auto_reset=False, fill_last_batch=True, dynamic_shape=False, last_batch_padded=False, device="cpu", device_id=0)
 
- Public Member Functions inherited from rocAL_pybind.amd.rocal.plugin.tf.ROCALGenericIteratorDetection
def next (self)
 
def __next__ (self)
 
def reset (self)
 
def __iter__ (self)
 
def __del__ (self)
 

Additional Inherited Members

- Data Fields inherited from rocAL_pybind.amd.rocal.plugin.tf.ROCALGenericIteratorDetection
 loader
 
 tensor_format
 
 multiplier
 
 offset
 
 device
 
 device_id
 
 reverse_channels
 
 tensor_dtype
 
 bs
 
 output_list
 
 dimensions
 
 dtype
 
 output_tensor_list
 
 output
 
 bbox_list
 
 label_list
 
 labels
 
 bboxes
 
 img_size
 
 num_bboxes_list
 
 label_2d_numpy
 
 bb_2d_numpy
 
 target
 
 target1
 
 res
 
 l
 
 num_bboxes_arr
 

Detailed Description

ROCAL iterator for detection and classification tasks for TF reader.

It returns 2 or 3 outputs (data and label) or (data , bbox , labels) in the form of numpy or cupy arrays. Calling .. code-block:: python ROCALIterator(pipelines, size) is equivalent to calling .. code-block:: python ROCALGenericIteratorDetection(pipelines, ["data", "label"], size)

Parameters
pipelinesThe rocAL pipelines to use for processing data.
sizeThe size of the iterator.
auto_resetWhether to automatically reset the iterator after an epoch.
fill_last_batchWhether to fill the last batch with repeated data to match the batch size.
dynamic_shapeWhether the iterator supports dynamic shapes.
last_batch_paddedWhether the last batch should be padded to match the batch size.

The documentation for this class was generated from the following file:
  • /home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-rocal/checkouts/docs-6.1.1/rocAL_pybind/amd/rocal/plugin/tf.py